An integrated metric approach to flexible query answering for information system

An integrated metric-based strategy is presented for handling inadequate queries where qualifications may contain imprecise/vague term, nominal term, linguistic term and/or linguistic modifiers. Two unified models are proposed for handling composite qualifications which contain linguistic hedges (very) or (fairly) for query modification, and for studying the modification effect of (very) and (fairly) on simple qualifications. When the Euclidean metric is used as a measure of semantic proximity, the acceptance region of approximate answers obtained for the class of inadequate queries considered reduces to a simple search interval, without the need to compute item-target distances. This Euclidean solution can be implemented on-the-fly with no excessive metadata required. In query answering, users can have the flexibility to interpret various imprecise terms to suit their own information needs. They may repeatedly change the size of the tolerance number needed to search for relevant approximate answers for retrieval.